AIM: To evaluate PT derived fibrinogen determinations with reference to the Clauss fibrinogen assay using a Sysmex CA-6000 random access coagulation analyser. METHODS: Samples were analysed from normal subjects (n = 20), patients with renal or liver dysfunction (n = 25), critically ill patients (n = 25), patients receiving oral anticoagulant treatment (n = 50), and patients with a haemoglobinopathy (n = 127). Prothrombin times were performed using two thromboplastins: one derived from rabbit brain (Dade: Thromboplastin IS) and the other from recombinant human tissue factor (Dade: Innovin). Fibrinogen was assayed by the Clauss method using a commercial kit (Dade: Fibrinogen). RESULTS: The relation between Clauss fibrinogen and PT derived fibrinogen was found to be dependent on the patient's clinical group and source of the thromboplastin used. When the data from the above sample groups were pooled there was still a significant difference (p < 0.001) between Clauss fibrinogen and PT derived fibrinogen, irrespective of thromboplastin used. CONCLUSIONS: It is unsafe to use the PT derived fibrinogen for patient monitoring owing to non-uniform variability in response to clinical status and reagent employed; however, it may prove to be a useful screening test in a research environment for estimating fibrinogen levels among defined patient groups.